We develop predictive computational models for three problems in pathogen evolution. First, we use computational chemistry to analyze the evolutionary changes in molecular shape and charge that recurred during multiple zoonotic transfers of influenza A from birds to humans. The repeated evolutionary patterns from past zoonotic transfers provide the basis for computational models that predict the risk of future pandemic strains. Our computational models will be useful for screening viruses sampled during surveillance of recurring human outbreaks of influenza derived from avian hosts. Second, we develop computational models to predict the efficacy of alternative vaccination strategies when vaccinating repeatedly against a rapidly evolving pathogen. Smith et al. et al. recently suggested that repeat vaccination should avoid cross reactivity with past vaccines and, at the same time, target predicted epidemic strains of the pathogen. We extend this idea by using our previously developed method for predicting influenza evolution based on patterns of positive selection. We use existing influenza data and computational methods to test the hypothesis that improved vaccine efficacy can be achieved by vaccinating with the influenza strain that we predict will dominate a few epidemics into the future. We will also develop an expanded computational model to analyze vaccination strategies for other rapidly evolving pathogens. Third, we construct mathematical and computer models to study the conditions that maintain co-circulating pathogen strains and the conditions that favor one strain to replace another. We will evaluate whether interference competition between strains mediated by cross-reactive host immunity can explain the observed patterns of fluctuating influenza strains. Based on our influenza work, we will extend our models to other pathogens to predict when new strains may arise and outcompete current strains. Finally, we will collate and make publicly available nucleotide sequence and antigenicity data for influenza from the unpublished CDC and WHO archives. This research will help to identify emerging pathogen strains that pose significant risk of causing widespread human pandemics. This research will also expand the range of vaccination strategies for use against rapidly evolving pathogens such as influenza.